Multichannel Signal Separation Combining Directional Clustering and Nonnegative Matrix Factorization with Spectrogram Restoration

Abstract

In this paper, to address problems in multichannel music signal separation, we propose a new hybrid method that combines directional clustering and advanced nonnegative matrix factorization (NMF). The aims of multichannel music signal separation technology is to extract a specific target signal from observed multichannel signals that contain multiple instrumental sounds. In previous studies, various methods using NMF have been proposed, but many problems remain including poor separation accuracy and lack of robustness. To solve these problems, we propose a new supervised NMF (SNMF) with spectrogram restoration and a hybrid method that concatenates the proposed SNMF after directional clustering. Via the extrapolation of supervised spectral bases, the proposed SNMF attempts both target signal separation and reconstruction of the lost target components, which are generated by preceding directional clustering. In addition, we experimentally reveal the trade-off between separation and extrapolation abilities and propose a new scheme for adaptive divergence, where the optimal divergence can be automatically changed in each time frame according to the local spatial conditions. The results of an evaluation experiment show that our proposed hybrid method outperforms the conventional music signal separation methods.Index Terms-Multichannel signal separation, music signal processing, nonnegative matrix factorization (NMF), spectrogram restoration.

Determined Blind Source Separation Unifying Independent Vector Analysis and Nonnegative Matrix Factorization

“…IVA generally assumes the spherical Laplace distribution, which has the same variance for all frequency bins, as the source prior by setting G(y j,m ) = ∥y j,m ∥ 2 in (9). The proposed method using (23) assumes independent complex Gaussian distributions in each time-frequency slot [34], similarly to conventional MNMFs. This issue will be discussed in Sect.…”

“…Note that these normalizations never change the value of the cost function (23). The signal scale can be restored by applying a back-projection technique [13] after the optimization.…”

“…One effective way of achieving this is to utilize a sample sound of the target signal [21]- [23]. However, such supervision cannot be utilized in BSS.…”

Experimental analysis of optimal window length for independent low-rank matrix analysis

“…As another means of solving audio source separation, nonnegative matrix factorization (NMF) [13], [14] is widely used for both blind and informed source separation [15]- [20]. NMF is a parts-based low-rank decomposition and can extract some meaningful spectral patterns (bases) with their time-varying gains (activations) from an observed spectrogram.…”

Independent Low-Rank Matrix Analysis Based on Time-Variant Sub-Gaussian Source Model

“…We compared the separation performance of the proposed sub-Gaussian GGD-ILRMA (β = 4) with those of conventional IS-ILRMA [9] and GGD-ILRMA (β < 2) [16]. We artificially produced monaural dry music sources of four melody parts (melody 1: main melody, melody 2: counter melody, midrange, and bass) using Microsoft GS Wavetable Synth, where several musical instruments were chosen to play these melody parts [20], [21]. Six combinations of sources, Music 1-Music 6, were constructed by selecting typical combinations of instruments with different melody parts.…”